Report terminology

The Strobe Performance Profile reports use the terms supplied in the following sections to measure resource use.

Terms displayed on the performance profile reports

Session time	The actual duration of the measurement session, expressed in minutes and seconds to the nearest hundredth.
CPU time	The amount of CPU time consumed by tasks executing in the measured job step, exclusive of the measurement task itself, expressed in minutes and seconds to the nearest hundredth. Because you are usually billed for the CPU time used, it is considered the most important measure of system resource time.
EXEC time percent	The percentage of execution time during which the central processing subsystem (comprising one or more CPUs) was in use by application tasks executing within the measured job step. Strobe computes the value as the ratio of samples in which it observed one or more sites of execution to the total number of samples taken.
Stretch time	The estimated amount of time that the CPU was unavailable to process programs executing in the measured address space because of demands made by higher-priority address spaces and by service request block (SRB) processing time for all address spaces. Stretch time may be high when multiple logical partitions (LPARs) are executing on the same complex of processors. The value in the STRETCH TIME field reported on Measurement Session Data reports can vary widely, even when you run Strobe under what seem to be identical conditions. This may result from a variation in system load.
SRB time	The amount of time during which the system was executing service request blocks (SRBs), which represent high-priority service routines.
Wait time	The portion of run time during which no task within the measured address space was able to make use of the CPU time available to it. Wait time usually occurs when a step is waiting for I/O to complete, an operator to respond, a user to provide input, or data to come from an allied address space.
Total time and Solo time	Strobe shows percentages for both total and solo resource usage time. Total time is a percentage of time that represents all the time during which the activity used a resource. This percentage includes all CPU execution time and all wait time. Solo time is the percentage of total time during which Strobe detected one and only one activity for the measured job step. Strobe measures solo time to highlight instances of resource use that have a direct effect on the run time of a batch processing program. In the case of task activity, solo time is the percentage of time during which only one CPU serviced the measured job step, with no concurrent file access activities. For file access activities, Strobe reports both CPU and I/O time. CPU time is the percentage of time spent in operating system data management modules. The I/O time reported is the percentage of time a device serviced a file access activity. Therefore, solo time can be either CPU time or I/O facility time spent servicing a file access activity, so long as neither task activity nor any other file access activity was serviced within the address space during that time.
Run time	The amount of time during which Strobe measured. It includes CPU time, comprising one or more CPUs, and wait time.

Pseudo-Entities

During its execution, the program or subsystem spends some portion of CPU time in supervisory routines that service components of the operating system control program or subsystem. Instead of reporting in detail on each such routine, Strobe groups them by function and reports the CPU time spent performing each function as time spent within a Strobe pseudo-entity.

The name of a pseudo-entity always begins with the period character (.) and reflects the function performed collectively by the routines it represents. Strobe defines four classes of pseudo-entities:

• pseudo-activity—The pseudo-activity .FILEMGT groups file management operations such as opening and closing files.
• pseudo-module—The pseudo-module .SYSTEM groups all supervisory functions.
• pseudo-transaction—A pseudo-transaction combines online subsystem overhead functions that Strobe cannot assign to any specific transaction. For example, the pseudo-transaction .CICS includes all CICS supervisory functions that Strobe could not assign to a user-defined transaction.
• pseudo-section—A pseudo-section is a collection of either:
  
  • Dynamically linked load modules, such as .IOCS (data management modules), or
  • Compiler library subroutines that are control sections within the target program load module, such as .COBLIB (COBOL library routines).

Pseudo-sections are subsets of either a real load module or the pseudo-module .SYSTEM. The following figure shows a Program Usage By Procedure report that displays pseudo-sections. The report lists the module’s name or the SVC’s number, the control section name (if Strobe obtained one during sampling), and a brief description of its function. Normally, a report does not show detailed module or control section information for system routines. You can, however, obtain such a breakdown for selected modules by using the DETAIL parameter on the Strobe - Detail For a Performance Profile panel and supplying the module names to Strobe during reporting. Refer to the Using-Strobe-to-measure-online-applications-and-batch-programs for more detailed information.

See Understanding-program-structure-and-Strobe-Pseudo-Entities for a description of all Strobe pseudo-entities.

Control Section Subreport—Pseudo-Sections